HAL Id: hal-03183442
https://hal.archives-ouvertes.fr/hal-03183442
Submitted on 29 Mar 2021
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
Segmental Dynamics of Poly(ethylene oxide) Chains in a Model Polymer/Clay Intercalated Phase: Solid-State
NMR Investigation
Cédric Lorthioir, Françoise Lauprêtre, Jérémie Soulestin, Jean Marc Lefebvre
To cite this version:
Cédric Lorthioir, Françoise Lauprêtre, Jérémie Soulestin, Jean Marc Lefebvre. Segmental Dynamics of Poly(ethylene oxide) Chains in a Model Polymer/Clay Intercalated Phase: Solid-State NMR Investi- gation. Macromolecules, American Chemical Society, 2009, 42 (1), pp.218-230. �10.1021/ma801909s�.
�hal-03183442�
1
Segmental Dynamics of Poly(ethylene oxide) Chains in a Model Polymer/Clay Intercalated Phase:
Solid-State NMR Investigation
Cédric Lorthioir,*,† Françoise Lauprêtre,† Jérémie Soulestin,‡ Jean-Marc Lefebvre‡
(†) Equipe "Systèmes Polymères Complexes",
Institut de Chimie et des Matériaux Paris-Est (UMR 7182 CNRS / Université Paris XII),
2-8 rue Henri Dunant, 94320 Thiais, France
(‡) Laboratoire de Structure et Propriétés de l’Etat Solide (UMR 8008 CNRS / Université Lille I),
Université des Sciences et Technologie de Lille,
Bâtiment C6, 59655 Villeneuve d’Ascq Cedex, France
2 AUTHOR EMAIL ADDRESS.
[email protected] (C. Lorthioir) [email protected] (F. Lauprêtre) [email protected] (J. Soulestin)
[email protected] (J.-M. Lefebvre)
RECEIVED DATE (to be automatically inserted after your manuscript is accepted if required according to the journal that you are submitting your paper to)
TITLE RUNNING HEAD.
Segmental Dynamics in a Model Intercalated Phase.
CORRESPONDING AUTHOR FOOTNOTE.
C. Lorthioir, Tel.: +33 1 49 78 13 08, Fax: +33 1 49 78 12 08, E-mail: [email protected]
3 ABSTRACT
A model poly(ethylene oxide) (PEO)/laponite hybrid material, characterized by a high silicate content, was used to probe the dynamical behaviour of polymer chains at the surface with clay platelets. Such a system mimics the intercalated phases that may occur in polymer/clay nanocomposites with usual silicate amounts of 5 wt %. The segmental motions underlying the -relaxation of fully amorphous PEO chains confined within the nanometer-thick laponite galleries were monitored over the tens of microseconds time scale by means of 13C and 1H solid-state NMR. A significant slowing down of these motions was mostly observed, as compared to the local dynamics in the amorphous phase of neat PEO.
Strong dynamical heterogeneities among the intercalated PEO monomer units remain even at room temperature, i.e. more than 50 K above the temperature at which the frequency of the segmental motions displayed by a significant part of the PEO chain segments gets above 52 kHz. Such heterogeneities are related to a pronounced extension of the -relaxation process towards the low frequency side. The slowing down of the PEO segmental motions was assigned to ion-dipole interactions between the PEO oxygen atoms and the Na+ counterions located in the laponite galleries. The domains formed by PEO monomer units characterized by a reduced segmental mobility were found to display rather long lifetime, about 13 ms at room temperature.
4 MANUSCRIPT TEXT
1. Introduction.
Polymer-based nanocomposites are of great current interest since the incorporation of inorganic nanosized fillers may, in some cases, lead to a strong enhancement of several macroscopic properties of the pristine polymer matrix.1,2 Polymer/clay structures are one of the most widely studied classes of organic/inorganic nanocomposites. Significant improvements of the mechanical, thermal and barrier properties of such materials with respect to the neat polymer matrix on the one hand and microcomposites prepared with the same matrix on the other hand have been reported in the literature.3,4 As far as the mechanical behaviour is concerned, one of the key parameter that must be optimized is the contact area between the polymer matrix and the filler particles. This experimental trend suggests that the behaviour of polymer chains at the interface with the clay platelets plays an important role in the mechanical performances displayed by polymer/clay nanocomposites.
Several theoretical and simulations-based approaches were developed in order to get a fine description of the influence of the clay platelets on both conformational and dynamical properties of the surrounding polymer chains.5-10 The comparison of these works with experimental results are of a great interest and promising investigations were carried out along this line.6-8 However, from an experimental point of view, several difficulties make the study of the interfacial phenomena in polymer/clay nanocomposites a rather challenging problem.
First, various morphologies may be obtained by blending polymer chains with clay platelets, depending on the preparation conditions and the nature of both blend components. The bulk polymer/clay materials may contain tactoids dispersed in the polymer matrix, intercalated phases, exfoliated phases or, more usually, a mixture of these different morphological structures.11 The effect of the clay layers on the static and dynamic behaviour of the surrounding polymer chains depends on the local polymer/clay organization. Indeed, the polymer chain radius of gyration Rg is limited by the thickness of the clay interlayer galleries in the case of intercalated phases, i.e. about 1.0 nm, while for
5 exfoliated phases, Rg is limited by the average interlayer distance, which may amount to at least a few tens of nanometers. Of course, the value of this distance depends on the clay content in the nanocomposite. As a result, a less severe geometrical chain confinement should be involved in the exfoliated phases and a lower local constraint on the polymer chains is expected. This difference is one of several features that should make the chain behaviour at the interfaces with the clay layers distinct between intercalated and exfoliated phases. Therefore, a fine description of the polymer chains in the vicinity of the clay platelets should ideally be carried out on isolated structural morphologies. These ones cannot be always easily obtained. One way to circumvent this first difficulty is to use a high clay content, which may lead, in some cases, to neat intercalated phases.12-14 The preparation of pure exfoliated phases is more complex and several strategies have been recently developed for this purpose.15,16
The second difficulty is related to the experimental approaches available to probe interfacial phenomena, from a dynamical point of view. Broadband dielectric relaxation spectroscopy has proved to be a very efficient approach to investigate the dynamics in heterogeneous polymer systems, covering a wide frequency range. In particular, this relaxation technique was used to probe the segmental dynamics as well as the fluctuations of the end-to-end vector for polymers with molecular dipole vectors oriented parallel to the main chain, in polymer/clay nanocomposites.13,14,17-19 The effect of the clay platelets on the segmental and normal mode of the polymer chains was considered. These studies were mostly carried out on polymer matrices filled with rather modest clay contents, typically lower than 5- 10 wt %, since this is the typical value required to get a significant enhancement of their macroscopic properties.17,18 However, the nanocomposite materials considered in these studies usually consisted in a mixture of different polymer/clay morphologies. In addition, in some cases, the conductivity contribution to the dielectric permittivity increases with the clay content and an additional Maxwell/Wigner/Sillars process occurs.17,18 Both contributions may superimpose with the relaxation processes related to polymer chains and the data analysis is not straightforward. For these reasons, nanocomposites characterized by a high clay content are not necessarily well-suited to probe the chain dynamics using dielectric relaxation spectroscopy measurements, even though such a strategy may
6 allow to get a pure morphological phase (intercalated phase). At this stage, it should be noted that this is not the case for all the clays. Indeed, Schwartz et al. took advantage of broadband dielectric relaxation spectroscopy to investigate both segmental and end-to-end vector dynamics of poly(propylene glycol) (PPG) chains within samples characterized by high vermiculite clay contents, high enough to get a neat intercalated phase.14 In this work, the contribution of both conductivity and interfacial polarization process was limited and both PPG segmental and normal relaxation peaks could be directly observed and characterized on the basis of isothermal measurements of the dielectric loss ’’(). No significant change in the -relaxation process was detected, in contrast to the normal mode. Similarly, Anastasiadis et al. considered poly(methylphenylsiloxane) (PMPS) oligomers intercalated in the galleries of two different organically-modified clays, montmorillonite and hectorite.13 Again, high values of the clay weight fraction, above 70 wt %, were used. In contrast with the results observed on PPG, the PMPS - relaxation process measured under isothermal conditions displays a strong shift towards high frequencies, as compared to the one observed on neat PMPS.
Solid-state nuclear magnetic resonance (NMR) spectroscopy is a powerful technique that has been successfully used in the last few years to describe various structural and dynamical features of polymer/clay nanocomposites, at the molecular length scale. Most of the polymer/clay nanocomposites described in the literature are based on naturally occurring layered silicates, which may contain paramagnetic species in more or less important quantities. In montmorillonite platelets for instance, the octahedral sites may be occupied by paramagnetic Fe3+ ions, instead of Al3+. VanderHart et al. took advantage of the paramagnetic contribution to the 1H relaxation properties in order to characterize the dispersion of the clay platelets in polystyrene or nylon-6 matrices and to investigate the nature of the nylon-6 crystallites (-phase and/or -phase) close to the silicate layers.20-22 However, most NMR studies were dedicated to synthetic clays or naturally occurring clays with a low content of paramagnetic species. The conformational behaviour of poly(ethylene oxide) (PEO) chains in intercalated phases was probed by 1D 13C NMR experiments23 and, more recently, a more precise description was obtained by means of 2D double-quantum experiments carried out on 13C-13C labelled
7 PEO monomer units.24 Solid-state NMR also served the study of the different dynamical processes involved in these nanocomposites. 7Li and 23Na NMR line shape analysis and spin-lattice relaxation time (T1) measurements were performed to monitor the dynamics of the counterions located in the clay interlayer galleries.25-28 From another point of view, the molecular motions of polymer chains were probed within nanocomposites based on PEO chains which display strong attractive interactions with the silicate layer surfaces.25,26,29-31 The case of polymers that do not display any propensity to intercalate the clay interlayer galleries, such as polystyrene (PS), was also considered.12,31 This is a more complex situation from a dynamical point of view, since a third component, quaternary ammonium salt surfactants usually, has to be introduced to enable the polymer chain intercalation. At this stage, it is worth noting that most of the NMR studies dedicated to the polymer chain dynamics within clay-based nanocomposites involved NMR experiments sensitive to the molecular motions occurring on the microsecond time scale. Investigation of the chain dynamics on the millisecond time scale was carried out recently on PEO and PS-based nanocomposites using two-dimensional 2H exchange NMR experiments.31 Moreover, the segmental dynamics in the shorter time scale range of 10-11 – 10-7 s was recently addressed in the case of PEO/synthetic fluoromica intercalated phases by means of electron spin resonance.32 For this purpose, nitroxide labels located at one of the PEO chain ends were introduced.
The references to NMR investigations mentioned previously concern polymer/clay nanocomposites with high clay contents. Interestingly, Schmidt-Rohr et al. proposed powerful 2D 29Si-
1H, 1H-1H and 1H-13C NMR approaches to extend the study of the chain dynamics at the interface with the clay layers, to nanocomposites composed of lower amounts of clay platelets.30 These experiments enable to describe the nature and the mobility of the species located in the vicinity of the inorganic layers. Such 2D NMR methods were successfully applied to nanocomposites with rather complex polymer matrices such as block copolymers33 or miscible polymer blends.34
Investigations of the polymer chain dynamics within clay galleries should serve to a better understanding and modelling of the mechanical properties displayed by polymer/clay nanocomposites as well as the dynamics in ultrathin polymer films. Indeed, the intercalated phases offer a 2D
8 confinement geometry where the polymer chains form (ultra)thin films, characterized by a rather well- defined confining length of about 1 nm. This characteristic length is much smaller than the ones that can be achieved using usual confining matrices35 and should amplify the confinement effects on the chain dynamics. In this respect, one may expect the geometric confinement imposed by the clay layers on the polymer chains to result in segmental motions faster than in the bulk matrix, as reported in previous investigations of PEO/clay intercalated phases for instance.19,26,29,31,32 In the case of PEO, several features were proposed to account for this confinement effect on the segmental dynamics. First, the local density was found to be lower for PEO chains confined by phyllosilicate platelets than for amorphous PEO in the bulk state.32 Second, as the temperature tends to the glass transition temperature of bulk PEO, the PEO segmental motions in a confined space should mainly involve rather localized, weakly cooperative motions.19,32 This limitation may lead to a correlation time value shorter than the characteristic correlation time related to the -relaxation of bulk PEO, at the same temperature. Lastly, the possible decrease of the chain entanglement density in a 1D-confined space by comparison with the bulk case could also explain the enhanced segmental dynamics.32
Besides, the dynamical behaviour of polymer chains in thin films and the characterization of its heterogeneity are yet open topics. The question of the glass transition phenomenon in thin polymer films was recently addressed, from a theoretical point of view, through several different models, which were mostly dedicated to Van der Waals liquids.36,37 However, to our knowledge, these approaches were not yet extended to the case of polymers involved in other kinds of interactions, such as hydrogen bonds for instance. Experimental data in this field may fruitfully serve the development of such theoretical models.
With these features in mind, we have considered polymer/clay nanocomposites, poly(ethylene oxide) (PEO)/laponite namely, characterized by high clay contents (50-70 wt %). Poly(ethylene oxide) displays attractive interactions with the large faces of the clay platelets and does not require the use of surfactants, such as quaternary ammonium salt surfactants, to intercalate the silicate galleries. Such low- molecular-weight molecules may indeed affect the segmental motions of the polymer chains and make the analysis of the dynamical behaviour more complex. Laponite was chosen as the clay component for
9 two reasons. First, this phyllosilicate consists in platelets displaying a discoid shape, with well-defined dimensions: the clay layer thickness is of about 0.92 nm while its diameter amounts to 25 nm. These dimensions display a rather limited distribution and the disc diameter is at least less polydisperse than the lateral size of natural clay layers. Moreover, naturally occurring clays may contain paramagnetic species, such as Fe3+ ions in the case of montmorillonite, the concentration of which depends on their nature and their origin. Even though the effect of such paramagnetic species on the relaxation behaviour of the surrounding protons may be fruitfully used to probe the clay platelet dispersion within the polymer matrix,20-22 it hampers the detailed analysis of the 13C or 1H NMR line shape of the chain segments in the vicinity of the inorganic layers. Therefore, for these reasons, laponite, which is a synthetic clay, was considered in the present work. Lastly, even though a few weight percent of clay is enough to induce a significant increase of the Young modulus, we focused on PEO/laponite nanocomposites characterized by high filler loadings, above 50 wt % more precisely. Our aim is to get a neat intercalated phase, where all the PEO chains are confined within the interlayer galleries. In such a situation, the PEO chains should be fully amorphous38 and the PEO segmental dynamics in this amorphous phase is not affected by the occurrence of PEO crystallites: under these conditions, the influence of the clay platelets on the surrounding PEO chain segments can directly be probed. Besides, when the clay content is high enough, all the PEO monomer units are necessarily in the vicinity of a clay layer, so that 1D 13C and 1H NMR experiments allow to investigate, in a straightforward way, the interfacial phenomena occurring at the laponite disc surfaces. As a result, the hybrid materials considered herein should be viewed as model systems, aimed at getting a better understanding of the chain dynamics in polymer/clay nanocomposites of applicative interests, i.e. characterized by clay contents of about 1-5 wt %.
In a first step, the fraction of laponite required to avoid PEO crystallization will be estimated and the bulk organization of these hybrid materials in the solid state will be determined using differential scanning calorimetry (DSC) and X-ray scattering experiments. In a second step, we will investigate a PEO/laponite model system corresponding to fully amorphous PEO chains confined within the nanometer-thick galleries formed by the laponite platelets. We will address questions concerning the
10 respective roles of chain confinement effects, which are expected to induce faster segmental motions than in the amorphous phase of the bulk PEO matrix, and attractive interactions of PEO chain segments with the hydrophilic laponite surfaces or complexation phenomena with the Na+ cations located within the Van der Waals gaps, which should contribute to decrease the PEO segment motions. Since both confinement and polymer/clay surface or polymer/Na+ cation attractive interactions should result in large variations of the PEO chain segment mobility and opposite effects, the PEO segmental motions in the neat PEO/laponite intercalated phase will be monitored by taking advantage of the selectivity of solid-state NMR. 13C and 1H NMR experiments will be used to investigate the influence of the clay layers on the dynamical behaviour of the surrounding PEO monomer units over the microsecond time scale. Special attention will be paid to the study of the heterogeneous character displayed by the segmental dynamics of the PEO chains.
2. Experimental Section.
2.1. Materials and Preparation.
Poly(ethylene oxide) (PEO, Aldrich), characterized by gel permeation chromatography, has the following molecular characteristics: weight-average molecular weight Mw = 123000 g.mol-1, polydispersity index Iw = 4.6. The clay used in this work is Laponite RD (Laporte Industries Ltd), which is a 2:1 phyllosilicate with the following chemical composition: Na+0.7[(Mg5.5Li0.3)Si8O20(OH)4]-0.7. This is a synthetic clay and as such, it does not contain any paramagnetic Fe3+ species embedded in the clay platelets.
PEO/laponite nanocomposites were prepared using the exfoliation-adsorption technique. The detailed conditions of their elaboration are described elsewhere39 and only the main features are recalled here. Distilled water was used to form a suspension of laponite clay at a concentration of 1 wt %. The suspension was stirred for 24 h, with a control of both pH (maintained to 10) and ionic strength (NaCl concentration of 10-3 mol.l-1). PEO homopolymer was then added, followed by a 24 h stirring period.
The nanocomposites were then obtained by freeze-drying. The PEO/laponite materials considered in
11 this work were characterized by a clay content of 50 and 70 wt % and will be denoted as PEO/laponite(50/50) and PEO/laponite(30/70) in the following.
Small-angle X-ray scattering (SAXS) experiments carried out on these samples, also reported elsewhere,39 showed the intercalation of the PEO chains in the silicate galleries, with a interlayer distance increasing from 0.94 nm (neat laponite) to 1.86 nm for both PEO/laponite nanocomposites. The first-order Bragg diffraction peak, related to the clay platelet stacking, was found to be narrower in the nanocomposite than in neat laponite. This feature suggests an increase of the orientational order locally displayed by the clay layers when blended with the PEO chains and is consistent with previous SAXS studies.38
2.2. Solid-State NMR Experiments.
Solid-state NMR experiments were carried out on a Bruker DSX 300 NMR spectrometer operating under a static magnetic field H0 of 7.0 T, equipped with a double-resonance MAS probe (4 mm rotors). The high-resolution 13C NMR spectra were obtained using either 1H → 13C cross- polarization (CP) or 13C direct polarization (DP) experiments. For the CP-based experiments, a 1H 90
pulse width of 3.0 s and a recycle delay of 3 s were used. The 1H field during the 1H → 13C magnetization transfer was set to 83 kHz. For 13C direct polarization experiments, a 13C 90° pulse width of 3.2 s and a recycle delay ranging between 4 s and 20 s were selected. The 13C-signal detection was performed under CW high-power proton decoupling (DD), with a RF field strength of 52 kHz. For most of the 13C NMR experiments presented in this work, the sample spinning speed was set to 5 kHz.
Occasionally, variations of the 1H decoupling field strength (from 19 to 100 kHz) and the MAS spinning speed (from 2.5 kHz to 9.0 kHz) were used to probe their influence on the PEO 13C NMR line shape.
The 13C chemical shifts were calibrated relative to tetramethylsilane (TMS) by taking the 13C chemical shift of the carbonyl peak of -glycine as an external reference standard (176.03 ppm with respect to TMS). 1H NMR spectra under MAS conditions were recorded using a 1H 90° pulse width of 3.0 s, a recycle delay of 10 s and a spinning speed of 5 kHz. Spin-lattice relaxation times in the laboratory
12 frame, T1(1H), were measured using the inversion-recovery pulse sequence while spin-lattice relaxation times in the rotating frame, T1(1H), were determined with a 90°-spin lock experiment, the spin-locking field corresponding to 66 kHz. The 1H chemical shift values were referenced using the methyl protons of ethanol as an external reference standard (1.11 ppm with respect to TMS).
1D 13C and 1H MAS NMR experiments were performed at variable temperatures, ranging from - 70 °C up to +100 °C. The temperature calibration of the MAS probe used for these measurements was carried out by monitoring the 207Pb chemical shift of lead nitrate.40 Besides, the sample volume was restricted to the centre of the rotor in order to limit the effect of temperature gradients on 13C and 1H NMR line shapes.
2D 1H-1H correlation experiments were carried out at a MAS spinning speed of 9 kHz, with a 1H 90° pulse width of 2.8 s and a recycle delay of 3 s. A detailed description of this experiment, in the context of polymer/clay nanocomposites, can be found in reference 30. A frequency-switched Lee- Goldburg 1H-1H decoupling at 89 kHz, with a 1H resonance offset of +/-59 kHz, was used during the evolution step. A variable mixing time tm allowing 1H-driven spin-diffusion to take place was introduced before the detection of the 1H NMR signal. The tm values ranged between 5 s and 289 ms.
2.3. Differential Scanning Calorimetry (DSC).
Differential scanning calorimetry experiments were performed on a TA Instruments DSC 2910 calorimeter. All the samples were submitted to the same thermal treatment: they were first heated at 100
°C, let at this temperature for 5 min and cooled down to -20 °C at a cooling rate of 10 °C.min-1. Two consecutive heating and cooling scans from -20 °C to 100 °C at 10 °C.min-1 were then recorded. These two heating scans were found to be identical, suggesting that no sample degradation or no evolution in the polymer/clay solid-state organization (in the case of the nanocomposites) occurred during the DSC experiments.
The degree of crystallinity of the PEO matrix was determined assuming a value of 188.6 J.g-1 for the heat of melting per unit mass of a 100% crystalline PEO sample.41 In the case of the PEO/laponite
13 nanocomposites, the clay weight fraction was taken into account in the calculation of the degree of crystallinity.
3. Results.
3.1. Solid-State Organization of PEO/Laponite Nanocomposites.
Figure 1 shows the DSC thermograms obtained on the PEO/laponite(50/50) and PEO/laponite(30/70) nanocomposites, during the second heating scan. These samples were previously cooled from 100°C to -20°C, at a cooling rate of 10°C.min-1. The thermogram obtained on the corresponding PEO homopolymer with the same experimental conditions is also reported for the sake of comparison. Under the thermal history indicated previously, a melting peak Tm at 63°C with an enthalpy
Hm of 145.0 J.g-1 is observed on neat PEO, indicating that the PEO matrix is characterized by a degree of crystallinity of 77%. Besides, the Gibbs-Thomson equation allows to get an estimate of the crystalline lamellae thickness of about 4.2 nm.42,43 In the PEO/laponite(50/50) nanocomposite submitted to the same thermal conditions, a PEO melting peak is also observed (Figure 1a), but the PEO matrix displays a smaller amount of crystallites (degree of crystallinity of 16% against 77% for neat PEO) while the crystalline lamellae thickness is reduced to 2.7 nm. In contrast, in the PEO/laponite nanocomposite containing 70 wt % of laponite, no PEO melting peak is detected in Figure 1b, which indicates that the PEO chains within this nanocomposite are amorphous.
14
0 20 40 60 80
-3,0 -2,0 -1,0
Heat flo w (W .g
-1)
T (°C)
T (°C) Heat flow (W.g-1)
PEO homopolymer
0 15 30 45 60 75 90 -2.0
-1.5 -1.0 -0.5
0 20 40 60 80
-2,0 -1,0 0,0
Heat flo w (W .g
-1)
T (°C) (a)
(b)
Figure 1. DSC thermograms recorded on (a) the PEO/laponite(50/50) nanocomposite, (b) the PEO/laponite(30/70) nanocomposite and on neat PEO (inset) during the second heating scan; the heating/cooling rate was set to 10 °C.min-1. The samples were previously cooled down to -20 °C from the molten state, with a cooling rate of 10 °C.min-1.
These observations are consistent with WAXS measurements performed on the same samples (Figure S1, Supporting Information). Indeed, Bragg diffraction peaks, characteristics of the monoclinic crystalline phase displayed by bulk PEO, can be observed on the diffractogram recorded on the PEO/laponite(50/50) nanocomposite, but the corresponding peaks are much less intense and significantly broader than in neat PEO. Both features are in agreement with the decrease of both
15 cristallinity and crystalline lamellae thickness, respectively, deduced from the DSC measurements. As the laponite weight fraction is increased to 70 wt %, the Bragg diffraction peaks related to PEO crystallites cannot be detected any longer on the WAXS diffractogram. Again, this result is consistent with the DSC experiments. Thus, with 70 wt % of laponite, all the PEO chains are mostly intercalated within the clay galleries and constitute a model system of amorphous chains confined to a narrow slit with a thickness of 0.96 nm. In the following, the PEO segmental dynamics will be exclusively probed in the PEO/laponite(30/70) nanocomposite since the occurrence of PEO crystallites within the PEO/laponite(50/50) sample, even in small amount and even if these crystallites are rather small, may significantly affect the segmental dynamics of the PEO chain segments in the amorphous phase.
3.2. Carbon-13 NMR Line Shapes.
The 13C CP/MAS/DD NMR spectrum recorded on the neat PEO homopolymer at room temperature is reported in Figure 2a. The 1H → 13C contact time for this experiment was set to 1 ms. A single 13C NMR peak at 70.4 ppm with a strong asymmetry in the high frequency side is observed. The
13C NMR line shape of semi-crystalline PEO has already been studied in many details in the literature.44-46 Thus, in the following, only the features that are relevant for the present work will be reminded. The line observed in Figure 2a results from the overlap of two peaks: one located at about 71.0 ppm, related to the PEO monomer units in the amorphous regions, and one occurring about 1.0 ppm downfield, assigned to the PEO units involved in the crystallites.45,46 The degree of crystallinity of the PEO sample considered for the experiment shown in Figure 2a amounts to 77%, as determined by DSC. Despite this rather high degree of crystallinity and the higher 1H → 13C CP transfer efficiency in crystalline regions (compared to PEO amorphous zones, at room temperature), the contribution from the crystalline PEO monomer units is not detected as a clearly resolved peak, due to its significant spectral width. This feature was investigated thoroughly in reference 44. One mechanism proposed to account for this result is based on the proton relaxation behaviour in the rotating frame, which is relevant during the 1H → 13C CP transfer. Indeed, at this temperature, the T1(1H) relaxation signal measured on the neat
16 PEO homopolymer considered in the present study is not monoexponential. The faster (longer) apparent T1(1H) component has been assigned to a dominant amount of crystalline (amorphous) domains.45,46 In a 13C CP/MAS/DD NMR experiment, during the 1H → 13C CP transfer, this T1(1H) relaxation feature tends to lower the amplitude of the 13C peak of crystalline PEO monomer units compared to the spectral contribution of the amorphous PEO carbons.
55 60
65 70
75 80
85
(ppm) (a)
(b)
(c)
Figure 2. 13C CP/MAS/DD NMR spectra of: (a) neat PEO and (b) PEO/laponite(30/70) nanocomposite;
the 1H → 13C CP contact time was set to 1 ms. (c) 13C DP/MAS/DD spectrum obtained on the PEO/laponite(30/70) nanocomposite; a recycle delay of 20 s was used for this experiment. All these spectra were recorded at room temperature.
17 Figure 2b depicts the 13C CP/MAS/DD NMR spectrum obtained with a contact time of 1 ms on the PEO/laponite(30/70) nanocomposite, at room temperature. The 13C NMR peak observed under these experimental conditions occurs at about 70.2 ppm and thus displays a weak upfield shift compared to the contribution of amorphous PEO carbons in the semi-crystalline PEO matrix described above. A deeper analysis of the 13C chemical shift displayed by the PEO carbons in the intercalated phase will be carried out in the following of this work. Besides, interestingly, the 13C NMR peak observed in the nanocomposite is not symmetric and displays a pronounced asymmetry in the low frequency side. This feature cannot be assigned to the contribution of PEO monomer units involved in crystalline regions, in contrast to the asymmetry observed under the same experimental conditions on the 13C NMR line of neat PEO. Indeed, both DSC and WAXS experiments presented above have shown that due to the high clay content used, the PEO chains within the nanocomposite are fully amorphous. In addition, the frequency location of the asymmetry observed in Figure 2b is not compatible with the 13C chemical shift of crystalline PEO carbons (Figure 2a). Additional 13C NMR experiments have been carried out to get a deeper understanding of the PEO 13C NMR line shape in the nanocomposite.
The 13C CP/MAS/DD NMR spectrum of the PEO/laponite intercalated phase may be compared to the 13C NMR spectrum obtained using a direct polarization (DP) experiment, reported in Figure 2c. In the DP/MAS/DD spectrum, the chemical shift of the peak related to the PEO units displays a slight upfield shift (less than 0.1 ppm) compared to the PEO contribution observed on the CP/MAS/DD spectrum. Moreover, the PEO 13C peak is much narrower than in the 13C CP/MAS/DD NMR spectrum shown in Figure 2b (88 Hz against 126 Hz) and the asymmetry is less pronounced. At this stage, it is of interest to recall that under the direct polarization conditions, all the PEO carbons equally contribute to the 13C NMR line. In contrast, in 1H → 13C CP-based experiments, the contribution of a given carbon to a 13C NMR peak depends on the strength of the heteronuclear 1H-13C dipolar couplings, but also on the spin-lattice relaxation time in the rotating frame T1(1H) involved by the protons surrounding this carbon. Therefore, in the PEO/laponite(30/70) nanocomposite, the difference in the 13C NMR line shape
18 related to the PEO carbons observed between DP/MAS/DD and CP/MAS/DD spectra may be explained by a strong distribution of the 1H-13C dipolar couplings occurring within the intercalated PEO monomer units and/or a non-uniform T1(1H) relaxation behaviour among the PEO protons. PEO carbons displaying a significant distribution of 1H-13C dipolar couplings would necessarily result from a pronounced heterogeneity in the segmental motions of the PEO chains confined between the laponite discs.
At this stage, it is of interest to consider the spin-lattice relaxation of the protons in the PEO/laponite(30/70) nanocomposite, in both laboratory (T1(1H)) and rotating (T1(1H)) frames. At a MAS spinning speed of 5 kHz, the 1H NMR peak at 3.6 ppm, related to the PEO protons, is rather well- separated from the one assigned to the laponite OH protons, located at 0.2 ppm. Thus, the 1H spin-lattice relaxation signals, reported in Figure 3, were measured under MAS conditions, using the height of both PEO and laponite 1H NMR peaks. Considering the height, instead of the area under the NMR peaks, enables to minimize the effect of their overlap on the 1H relaxation signals. As can be seen in Figure 3a, the T1(1H) relaxation function determined on both PEO and laponite 1H peaks are identical and can be described by a single-exponential decay, with T1(1H) = 0.25 s. This result indicates that the 1H magnetization related to the PEO chains fully equilibrates with the one of the clay OH protons on the T1(1H) time scale. In contrast, the T1(1H) relaxation decay of the PEO protons does not superimpose with the one obtained for the clay protons, as shown in Figure 3b. Such a difference shows that complete equilibration of the proton magnetization related to PEO and laponite components does not occur on the T1(1H) time scale. In addition, the T1(1H) relaxation is not monoexponential, but can be described using two distinct T1(1H) components: T1,s(1H) = 1.5 ms and T1,l(1H) = 5.3 ms for the relaxation signal obtained using the height of the PEO 1H peak. The amplitude related to the faster T1(1H) relaxing component of the signal determined on the PEO 1H peak is of about 41%. This latter is much more intense than the peak of the laponite OH protons so that the partial overlap of both peaks is not sufficient to account for the biexponential feature of the T1(1H) relaxation function determined at 3.6 ppm.
19
0.0 0.2 0.4 0.6 0.8 1.0
0.01 0.1 1
[ I
0- I( ) ] / I
0T
1(
1H) relaxation delay (s)
(a)
(b)
0 2 4 6 8 10
0.1 1
I( ) / I
01
H spin-lock time (ms)
Figure 3. Proton spin-lattice relaxation (a) in the laboratory frame, (b) in the rotating frame, measured on both PEO/laponite(30/70) nanocomposite components: () PEO protons and () laponite OH protons. The relaxation decays were deduced from the height I of both PEO and clay proton peaks, observed under MAS conditions (spinning speed of 5 kHz). The data were normalized by I0, the proton magnetization at full equilibrium, under the static NMR magnetic field H0.
20 Non-uniform T1(1H) relaxation behaviour among the PEO monomer units may be induced by dynamical heterogeneities in the PEO segmental motions within the laponite galleries, leading to distinct contributions to the spectral density in the upper-kilohertz region (66 kHz, the intensity of the
1H spin-lock field used, expressed in frequency unit). However, the effect of 1H-driven spin-diffusion should also be taken into account to interpret the T1(1H) relaxation behaviour of the PEO protons. In particular, the longer T1(1H) decay may be related to 1H-driven spin-diffusion from the protons of the laponite platelets. Thus, at this stage, the origin of the non-uniform T1(1H) relaxation of the PEO protons cannot be unambiguously determined. However, one important point to go further in the analysis of the 13C CP/MAS/DD NMR spectrum of the nanocomposite is that the faster T1(1H) relaxing component is longer than the 1H → 13C contact time of 1 ms used for the spectrum reported in Figure 2b. Thus, after the CP step, the amplitude of the short (long) T1(1H) decay measured at 3.6 ppm is reduced to about 50% (83%) of its initial value and one might expect that the non-uniform T1(1H) decay of the PEO protons is not strong enough to account for the differences in the 13C NMR line shape detected between CP/MAS/DD and DP/MAS/DD spectra.
In order to confirm the previous assumption, 13C CP/MAS/DD NMR spectra were recorded on the intercalated PEO chains for various 1H → 13C contact times, tCP, and representative results are plotted in Figure 4. Let us first consider the results obtained using tCP values that are small compared to the faster T1(1H) relaxing component determined for the PEO protons, namely tCP = 20 and 100 s. For these contact time values, the PEO 13C line shape also displays strong differences with the one observed on the DP/MAS/DD spectrum of Figure 2c. This feature clearly evidences that the non-uniform T1(1H) relaxation of the PEO protons is not enough to account for the differences between 13C CP/MAS/DD and DP/MAS/DD spectra of the nanocomposite obtained at room temperature. As a consequence, in the PEO/laponite intercalated phase, it must exist a significant distribution of the 1H-13C dipolar couplings among the PEO chain segments, corresponding to a strong heterogeneity in the PEO segmental motions.
21
50 55
60 65
70 75
80 85
90
(ppm) (a)
(b) (c) (d) (e)
Figure 4. 13C CP/MAS/DD NMR spectra of the PEO/laponite(30/70) nanocomposite obtained for various values of the 1H → 13C contact time, tCP: (a) 20 s, (b) 100 s, (c) 400 s, (d) 2 ms and (e) 10 ms.
22 Considering now the variations of the 13C NMR line shape with the 1H → 13C contact time (Figure 4), one may note that for short tCP values (20 s for instance), the peak is rather broad (half-height line width 1/2 of 351 Hz) and symmetric. As tCP is increased, the 13C NMR peak gets narrower, the asymmetry is more and more pronounced and a low field shift of the peak is also observed. Figure 5 depicts the evolution of both line width at mid-height 1/2 and 13C chemical shift with the 1H → 13C contact time. Both parameters tend to a constant value for tCP higher than 2 ms. Therefore, the analysis of the variation of the PEO line shape in the nanocomposite with the 1H → 13C contact time enables to rationalize the asymmetric 13C line shape initially observed on the 13C CP/MAS/DD NMR spectrum obtained with a contact time of 1 ms (Figure 2b). The upfield asymmetry results from the superposition of different contributions related to PEO chain segments characterized by distinct 1H-13C dipolar couplings and so, distinct segmental mobility, as well as distinct chemical shifts: the lower the dipolar couplings (and thus the higher the segmental mobility), the higher the 13C chemical shift.
0.01 0.1 1 10
100 200 300 400
t
CP(ms)
1/2
(H z)
69.2 69.6 70.0 70.4
(pp m)
Figure 5. Variation of the half-height line width 1/2 () and the chemical shift () of the PEO 13C NMR peak with the 1H → 13C contact time tCP. These values were measured on the 13C NMR spectra of the nanocomposite, recorded through 1H → 13C CP/MAS/DD NMR experiments.
23 Having this interpretation in mind, one should expect the heterogeneous dynamical behaviour of the PEO chain segments, observed at room temperature, to disappear at sufficiently high temperature.
Thus, both 13C CP/MAS/DD and DP/MAS/DD NMR spectra were recorded on the PEO/laponite(30/70) nanocomposite, at 373 K. For the 13C CP/MAS/DD NMR experiment, a contact time of 1 ms was selected. At this temperature, both spectra are identical, as can be seen on Figure 6.
(a)
(b)
64 68
72 76
(ppm)
Figure 6. Comparison of the 13C NMR spectra obtained at 373 K on the PEO/laponite(30/70) nanocomposite using: (a) 1H → 13C CP/MAS/DD experiment, with a contact time tCP of 1 ms, (b) 13C DP/MAS/DD NMR experiment.
24 This result demonstrates that at 373 K, all the PEO chain segments display similar values of the
1H-13C dipolar couplings and thus, very close segmental mobilities. Increasing the temperature by about 80 K leads to the suppression of the pronounced heterogeneities in the PEO segmental motions detected at room temperature, as expected from our analysis of the data obtained at room temperature. It is worth remarking that, in the PEO/laponite(30/70) nanocomposite, the chemical shift of the PEO 13C NMR peak is the same at room temperature and at 373 K. Only the upfield asymmetry has disappeared at high temperature. This last feature is consistent with the dependence of the 13C CP/MAS/DD NMR spectra with the CP contact time, observed at room temperature.
Complementary information on the PEO segmental dynamics in the PEO/laponite intercalated phase can be derived by monitoring the 13C NMR line shape variation as a function of temperature.
Illustrative 13C DP/MAS/DD NMR spectra recorded on the PEO/laponite(30/70) nanocomposite for various temperatures between 335 K and 203 K are shown in Figure 7 and the corresponding temperature dependence of the half-height line width 1/2 is reported in Figure 8. For comparison, the evolution of 1/2 with the temperature for amorphous monomer units of neat PEO is also plotted in Figure 8. Though bulk PEO is a semi-crystalline matrix, the selective observation of the amorphous phase is achieved by taking advantage of the difference in the T1(13C) value displayed by both crystalline and amorphous phases.47 Thus, as T1(13C) is much higher for PEO crystalline carbons than for amorphous PEO monomer units, 13C DP/MAS/DD NMR spectra recorded with intermediate recycle delays enable to probe, selectively, the amorphous phase. Spectra obtained under these conditions (recycle delays of 0.6 s) were used to determine the 1/2(T) values reported in Figure 8.
Let us first consider the data related to the PEO/laponite intercalated phase. From 335 K down to 305 K, 1/2 remains nearly constant, with a value of 73 Hz at 305 K. More precisely, a very weak increase of the line width is observed as the temperature is decreased in this range, but it can be neglected compared to the pronounced line broadening occurring from 73 Hz at 305 K up to 472 Hz at 243 K. The half-height line width reaches a maximum at 243 K and a further temperature decrease
25 induces a line narrowing until 223 K. From 223 K to 203 K, which is the lowest temperature that was reached experimentally under the MAS conditions, 1/2 remains nearly unchanged.
55 60
65 70
75 80
85 (ppm)
203 K
243 K
263 K
283 K
325 K
Figure 7. 13C DP/MAS/DD NMR spectra obtained on the PEO/laponite(30/70) nanocomposite for temperatures ranging between 335 K and 203 K. The MAS spinning speed was set to 5 kHz and the intensity of the CW proton decoupling during the 13C-signal acquisition, to 52 kHz.
26
210 240 270 300 330
0 100 200 300 400 500
1/2(H z)
T (K)
Figure 8. Temperature dependence of the half-height line width 1/2 of the PEO 13C NMR peak measured on the DP/MAS/DD spectra obtained for the PEO/laponite(30/70) nanocomposite () and for neat PEO (). In the case of the neat semi-crystalline PEO matrix, a short recycle delay was used to probe selectively amorphous chain segments.
The global line width reduction from low (203 K) to high (335 K) temperatures may be assigned to the crossover from the glassy to the molten state of the PEO chains intercalated within the clay platelets (-relaxation). The same explanation holds for the data obtained on amorphous monomer units of bulk PEO. In the glassy state, the frozen chain conformations as well as the variations in the interchain packing induce a dispersion of the 13C chemical shift values related to the PEO carbons. Far above the glass transition, such a dispersion is strongly reduced due to a strong motional averaging effect. In particular, in the molten state, one significant contribution to the 13C line width may origin
27 from the local magnetic field inhomogeneities. In the present case, the contrast between the diamagnetic susceptibility of the clay layers and the one related to the PEO chains may contribute to NMR magnetic field inhomogeneities which could be experienced by some PEO carbons. Experimentally, the line width at the high temperature plateau of the 1/2(T) variation is more than two times higher for the PEO/laponite(30/70) nanocomposite than for amorphous chain segments of the neat PEO homopolymer.
Such a difference may result from the diamagnetic susceptibility contrast effects mentioned above as well as differences in the PEO segmental dynamics between both samples, remaining up to 335 K.
From another point of view, the additional line broadening detected around 243 K on Figure 8 may be ascribed to the modulation of the 1H-13C dipolar couplings by the molecular motions of the PEO chain segments in the intercalated phase.44,48 More precisely, as the characteristic frequency of the PEO C-H bond motions matches the intensity of the 1H decoupling field, expressed in frequency unit, interference effects between the motional modulation of the 1H-13C dipolar couplings and the coherent modulation of the 1H decoupling field lead to an increase of the line width of the PEO 13C NMR peak.
Another dynamic line broadening mechanism, related to the modulation of the chemical shift anisotropy of the PEO carbons, could be invoked to account for the line broadening around 243 K observed in Figure 8. Such a motional modulation may interfere with the coherent modulation induced by MAS, as soon as the characteristic frequency of the reorientational motions displayed by the PEO carbons is equal to the MAS spinning speed r.44,49 To check whether this line broadening mechanism is responsible for the temperature dependence of 1/2 around 243 K, a 13C DP/MAS/DD NMR spectrum was recorded on the PEO/laponite(30/70) nanocomposite at this temperature, using a different value of the MAS spinning speed: 2.5 kHz. No significant variation in the PEO 13C NMR line shape was detected when comparing this spectrum with the one obtained at the same temperature, with a spinning speed of 5 kHz. In particular, the half-height line width is very similar in both cases. In the so-called weak collision regime, the contribution to 1/2 induced by the motional modulation of the chemical shift anisotropy, denoted as 1/(T2), is expected to be proportional to r-2 in the case of slow motions, i.e. when the characteristic correlation time of the probed motions is much higher than r-1.44,49 The
